/Stable-Alignment

Efficient, Effective, and Stable alternative of RLHF. Code for the paper "Training Socially Aligned Language Models in Simulated Human Society".

Primary LanguagePythonOtherNOASSERTION

No issues in this repository yet.